[ghc-steering-committee] Base library organisation

Sun Jul 9 21:27:20 UTC 2023

This protocol is tricky to get right.  But I think that the ideas sketched
in Section 6.1 come close.  (For completeness, here is the current draft
plan.)
<https://github.com/Ericson2314/tech-proposals/blob/ghc-base-libraries/proposals/accepted/050-ghc-base-libraries.rst>

First, we should make explicit in every proposal what library changes (if
any) there are, and to which libraries.  All by itself that will be helpful.

Second. the ghc-experimental library provides authors with more options
than we had in the past:

   - They can propose changes to base, and accept the additional
   constraints of discussing with the CLC; including any changes that
   subsequently prove to be necessary or desirable
   - Or, in many cases, they can deliver their proposal through
   ghc-experimental, which CLC is quite content for the GHC CS to curate.

Third, we envisage that the shepherd will explicitly invite the CLC to
express a non-binding opinion about any proposal modifying base.   That
greatly lessens the risk of a proposal being accepted that the CLC is
unhappy with.  No one wins if that happens!

The "non-binding" bit is important: the devil is sometimes in the details.
But remember that acceptance of a GHC proposal is *also* non-binding.  We
explicitly say that if the implementation adds too much complexity, or
shows up new issues that were not discussed when the proposal was accepted,
then we may not land it.  Acceptance is a decision in principle.

(Side note.  I don't think we should be legalistic about this, but if CLC
does not voice any objection to changes to base explicitly called out in
the proposal, then it would be very unusual for them to subsequently reject
outright a CLC proposal to implement those same changes.  Let's jump that
bridge if we come to it; I don't think we will!)

So fourth, the protocol does envisage that, as part of the implementation
process, the implementor will make a CLC proposal.  It should be no more
than a copy/paste of the relevant section from the proposal, but it'll be
at the point of "I firmly intend to implement this proposal, so it is worth
requesting CLC bandwidth to discuss the detailed specifics of the API
changes".

This all makes sense to me.  I hope it does to you all, as members of the
GHC SC. Please say if not.  I'm keen to get this tied up so we can
publicise it.

I would be happy to elaborate Section 6.1, to make it more clear and
specific, if anyone has some draft text they would like to propose.

Simon

On Sun, 9 Jul 2023 at 05:53, Vladislav <vlad.z.4096 at gmail.com> wrote:

> This proposal has just popped up in another discussion, and its
> implications dawned on me.
>
> I see two major issues with the proposal in its current form:
>
> 1. Proposal authors shouldn't need to interact with two committees instead
> of one. It is generally fine to have coordination between committees, so I
> can imagine a system where GHC SC cannot approve a proposal on its own and
> must contact CLC when a proposal touches base. Proposal authors mustn't be
> bothered with the intricacies of this process, they need to write a single
> document, submit it on a single platform, and then either have it accepted
> or rejected, whether by GHC SC alone or by a joint GHC SC + CLC committee.
>
> 2. Proposal implementors need a single source of truth for the accepted
> changes. If the proposal document has been marked accepted, it means that
> it's ready for implementation. Now, I have no problem whatsoever with
> involving CLC before a proposal is accepted, but once it's done, it's done.
> It's quite important to be able to point potential contributors to a
> document that describes the changes we want to see in GHC and its
> libraries, and it can't be that parts of this document are actually
> accepted and parts of it are
> kind-of-accepted-but-non-binding-and-another-proposal-on-another-platform-is-pending.
>
> So I agree with the general idea that it's important to get CLC involved,
> but for the sake of proposal authors and proposal implementors, this should
> be a committee-to-committee interaction, not a person-to-two-committees
> interaction; and acceptance of a proposal needs to be a single atomic
> operation instead of having kind-of-accepted documents floating around in
> the ghc-proposals repo.
>
> Vlad
>
> On jeu., juin 29 2023 at 09:49:09 +01:00:00, Simon Peyton Jones <
> simon.peytonjones at gmail.com> wrote:
>
>   make the GHC proposal in conjunction with a CLC proposal and land them
>> in GHC together. Streamlining this process would be good
>>
>
> That is absolutely an option, yes.
>
> Simon
>
> On Thu, 29 Jun 2023 at 08:27, Simon Marlow <marlowsd at gmail.com> wrote:
>
>> The link in the email is broken BTW, is there any updated URL?
>>
>> Just based on the summary in the email, what immediately springs to mind
>> is that adding functions to ghc-experimental and then moving them to base
>> implies an extra migration that clients have to do in the future (would the
>> APIs in ghc-experimental be deprecated first? For how long? etc.). It
>> probably makes sense only to do this for additions of fairly large new
>> APIs, and for smaller changes we should probably shortcut the process and
>> make the GHC proposal in conjunction with a CLC proposal and land them in
>> GHC together. Streamlining this process would be good. Maybe the actual
>> proposal already says a lot of this?
>>
>> Cheers
>> Simon
>>
>>
>> On Thu, 15 Jun 2023 at 10:04, Simon Peyton Jones <
>> simon.peytonjones at gmail.com> wrote:
>>
>>> Dear GHC Steering Committee
>>>
>>> Over the last few weeks, Ben Gamari and I have been discussing with
>>> Andrew and Julian from the Core Libraries Committee how to make the Core
>>> Libraries Committee and the GHC developers work together more fluidly; and
>>> that includes the GHC Steering Committee.
>>>
>>> We now have a fairly well fleshed out proposal here.
>>> <https://github.com/Ericson2314/tech-proposals/blob/ghc-base-libraries/proposals/accepted/050-ghc-base-libraries.rst>
>>>
>>> I hope you like it.  As far as this committee is concerned there are two
>>> particular points of note
>>>
>>>    1. We propose a new package, *ghc-experimental*, which depends on
>>>    *base*.  Many GHC proposals involve defining new types and
>>>    functions.  The idea is that these would initially be in
>>>    *ghc-experimental*.  After they stabilise and become widely adopted,
>>>    the author (or anyone else) can make a CLC proposal to move them to
>>>    *base*, which has much stronger stability guarantees.
>>>    2. Section 5.1 suggests a mechanism to involve CLC members in
>>>    proposals that involve new functions and types, at an earlier stage.  Some
>>>    involve *changing *existing types and functions.  It is clearly
>>>    unproductive for us to debate such things at length, and only *then *to
>>>    land it on the CLC.
>>>
>>>    Section 5.1 also suggests that proposals should explicitly (in a
>>>    separate section) call out
>>>
>>>
>>>    - What new types and functions it defines
>>>    - What existing types and functions are changed.
>>>
>>> We should add that to our template.
>>>
>>> At the moment we are just sharing the proposal with relevant
>>> stakeholders (yourselves, CLC, stack folk, cabal folk etc), so that we can
>>> polish any rough edges before making it public.
>>>
>>> So, any views?  Personally I think this is a Big Step Forward.
>>>
>>> Simon
>>>
>>>
>>>
>>> _______________________________________________
>>> ghc-steering-committee mailing list
>>> ghc-steering-committee at haskell.org
>>> https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-steering-committee
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-steering-committee/attachments/20230709/4f1e9dc4/attachment-0001.html>